The spectral analysis of microwave-frequency signals consists in identifying the frequency components of an unknown signal, and optionally their respective amplitudes.
Commercially available spectrum analyzers intended for signal characterization operate in the following way: the input signal is demodulated by mixing it with the signal of a local oscillator. After low pass filtering, the components present in a narrow band about the frequency of the local oscillator are obtained. It is then necessary to scan the frequency of the local oscillator in order to obtain all the information on the input signal, in a large spectral window.
The main drawback of this solution is that it is intrinsically slow because of the scan. In addition, it requires complex electronics and is therefore costly and bulky.
Lastly, in the case of signals that vary as a function of time, the prior art does not allow all of the spectral window to be monitored at the same time (unless one analyzer is employed for each band). The probability of interception of the signals is at best a few %.
For applications requiring an interception probability as close as possible to 100%, an alternative solution is to directly digitize the input signal over the largest possible passband, then to apply Fourier transforms to the samples obtained. This solution has two limits:                The best input passband achievable with current analogue-digital converters is less than 10 GHz, and they also have dynamic ranges that are far lower than those required (mainly because of sampling jitter and conversion nonlinearities).        The rate at which the data must be transmitted and processed is considerable (several hundred Gb/s and several Tflops) leading to a prohibitively high price, power consumption and volume.        
Although the ability of processors to process and manage data delivered at high rates is improving very rapidly (Moore's law), the same does not go for the performance of very-wideband analogue-digital converters, which is limited by the fundamental properties of the electronic components employed.
One aim of the present invention is to mitigate the aforementioned drawbacks by providing a system and method for analyzing a spectrum allowing an “image” of the complete spectrum of an input signal to be obtained over a wide band, without scanning nor digitization or processing.